314 REPORTS ON THE STATE OF SCIENCE, ETC. 



ment and that the normal function of perception is to provide information 

 about that environment. So we set up three suitable objects to provide 

 stimuli. 



We must be able to adjust at least one of the objects with respect to the 

 perceptible property in such a way that the stimulus we receive from it can 

 be varied in magnitude without change of quality. In the visual case, to 

 which I will confine the remainder of the description, we must be able to 

 alter the brightness of the object without changing its colour. There are 

 numerous practical methods of doing this. Let us denote the objects by 

 A, B and C, and suppose that C is brighter than A, and that B is brighter 

 than A but less bright than C. The pair of unequally bright objects A and B 

 present us with a directly perceived phenomenal relation. So also does the 

 pair B and C. We must not confuse these directly perceived phenomenal 

 relations with the quantitative relations between the measured intensities 

 of the stimuli. In order to know anything about the latter relations we have 

 to measure the stimulus values by appropriate methods, but in order to 

 perceive the phenomenal relations we have only to look at the objects. The 

 perceivable phenomenal relations would still be just what they are if we had 

 never formulated a scale of measurement for brightness as a photometric 

 magnitude, in which case there would be no quantitative relations between 

 the stimuli. So we cannot describe phenomenal relations, as intuitively 

 perceived, in terms of quantitative concepts. They are simply those aspects 

 of the objective world, as directly perceived, in virtue of which we 

 differentiate objects and groupings of objects from each other. Returning 

 to our experiment, we perceive the phenomenal relations between objects 

 A and B and between B and C, which we will denote hy A.B and B . C, the 

 notation conveying no implications about the nature of either relation. 

 We also notice a relation betzoeen these relations. We find that as we vary 

 the brightness of B between those of A and C not merely do the individual 

 relations A.B and B .C change but so also does this cross relation. We are 

 not required to describe (nor indeed are we able to describe) how it changes. 

 To say that the perceived interval AB is greater, equal to, or less than the 

 perceived interval BC is merely to use words before we know what they mean 

 in the particular application. The relation changes in some respect which 

 is directly apprehensible to us : that is all we can say. In general the rela- 

 tion is not one that we recognise as having any special significance, but as 

 we go on adjusting the brightness of B to various values we find that there 

 is one value, and only one, for which the relation we are now speaking of, 

 the relation between the relations A . B and B .C, is such that we recognise 

 those relations as having sameness in some respect which has a unique 

 significance in perceptual experience. 



This is the criterion we have been trying to satisfy : the experiment is 

 finished except for the measurement of the stimulus values, i.e. the actual 

 brightnesses oi A,B and C, by the appropriate photometric methods. The 

 experimental psychologist may object that to look for a recognisable relation 

 between undefined relations A . B and B .C is not what the observer is asked 

 to do, and there is therefore no reason for supposing that this is what he 

 does. It is the only thing he can do. The instructions usually given him, 

 as already stated, are couched in terms of operations for which no meaning 

 has been defined. Whether or not he thinks he understands them and 

 believes himself to be carrying them out is of no consequence. He cannot 

 in fact carry them out. All he really gathers from the instructions is that, 

 there is some unique perceptual criterion by which the stimuli may be 

 arranged, and assumes when he has found such a criterion that it is the one 

 the instructor means. Thus, although in many psycho-physical experi- 



